Add core copy detection algorithm This adds findcopies, which detects merge-relevant copies between files in a pair of manifests back to the merge ancestor. While the merge code invokes the copy detection routine, it does not yet use the result.

context: change filectx repr to use @ Avoids confusion with rev:nodeid syntax used elsewhere

context: add __repr__ methods

filectx: lazy linkrev usage

restore filectx.changectx() method

fix newline in error message

add comments in cachefunc

filectx.annotate: return filectx for each line instead of rev

Add cachefunc to abstract function call cache

Make filectx lazier - some users never use filenode

context: handle fileid or changeid == 0

merge with crew

mq: Add --rev argument to qimport, to adopt existing changesets.

manifest.py: remove unnecessary method

Switch revlog.ancestor to use revisions rather than nodeids

mpatch: Fix for malloc corner case on AIX

Merge with crew

tests: add new test; test-mq-qrefresh

Abstract ancestor algorithm into generic function Make depth calculation non-recursive Add simple shortcut for linear ancestry Convert context to use ancestor function make memoized parents function Convert revlog to use ancestor function

Merge with crew

qimport: rename patch to patchname to avoid shadowing module

Move defaultrev into changectx This also causes tag on a repository with no working directory to default to tip.

fix warnings spotted by pychecker

hgweb: fix errors spotted by pychecker

extdiff: use the default option only if the default program is used

Merge with mainline

Update [extdiff] configuration sample for vimdiff, taking acount cmd.xxx and opts.xxx configurations item.

filectx: add rename-aware ancestor algorithm This code works but may trigger recursion depth issues

changectx: add ancestor function

filectx: allow passing filelog in init to avoid opening new filelogs

filelog: make metadata method private

filectx: add rename traversal for parents()